Discussion Feed

Question   What is the best way to setup Alerts for App Agent Status and Machine Agent Status.     When the Agent is turned off  the Agent Status stops reporting any values so what is the best approach for setting up a Health Rule?    Answer Solution 1 The machine agent posts a metric you can use:   Application Infrastructure Performance|FrontEnd|Agent| Machine|Availability   Solution 2 You can also use the App Agent Status for app agent availability. But note in addition to the app pool recycles that take place by default every 29 hours, there is also an app pool idle time out that occurs if there’s no traffic on the app after 20 minutes. If that idle occurs, then the agent will be unloaded will no longer report and trigger this alert and look like a false alert (App|Availability = 0). You can address this by turning that setting off by setting timeout value to 0.     You can read about this here under the section Idle Time Out. ... View more

As of 4.0, license usage data can be viewed  in the license screen from the controller UI. See License Information for details.   SaaS users can view license usage information from their account subscription page on www.appdynamics.com.     The subscription page shows shows data for the current month with the most used product first. For each product area, the X-axis represents time in hours and Y-axis represents the units used for that area. For example, for APM it shows the agent license units used, for Transaction Analytics it shows BT events, for Mobile EUM active users, and so on.        By hovering over the graph, you can see how much is used for that hour of the day.  For APM products with either fixed or overage models (except variable model), an additional line is shown, which represents the number of units provisioned.     ... View more

Issue Java agent is unable to connect to the controller due to a certificate chaining error. This might be seen with IBM WebSphere.    Errors similar to the following:   [AD Thread Pool-Global0] 10 Jun 2015 20:12:52,848 WARN XMLConfigManager - Certificate chain validation failed com.ibm.jsse2.util.j: PKIX path building failed: java.security.cert.CertPathBuilderException: PKIXCertPathBuilderImpl could not build a valid CertPath.; internal cause is: java.security.cert.CertPathValidatorException: The certificate issued by CN=DigiCert Global Root CA, OU= www.digicert.com<http://www.digicert.com >, O=DigiCert Inc, C=US is not trusted; internal cause is: java.security.cert.CertPathValidatorException: Certificate chaining error attempting validation. [AD Thread Pool-Global0] 10 Jun 2015 20:12:52,849 ERROR ConfigurationChannel - Fatal transport error: com.ibm.jsse2.util.j: PKIX path building failed: java.security.cert.CertPathBuilderException: PKIXCertPathBuilderImpl could not build a valid CertPath.; internal cause is: java.security.cert.CertPathValidatorException: The certificate issued by CN=DigiCert Global Root CA, OU= www.digicert.com<http://www.digicert.com >, O=DigiCert Inc, C=US is not trusted; internal cause is: java.security.cert.CertPathValidatorException: Certificate chaining error [AD Thread Pool-Global0] 10 Jun 2015 20:12:52,849 WARN ConfigurationChannel - Could not connect to the controller/invalid response from controller, cannot get initialization information, controller host [stelo.saas.appdynamics.com< http://stelo.saas.appdynamics.com> ], port[443], exception [Fatal transport error: com.ibm.jsse2.util.j: PKIX path building failed: java.security.cert.CertPathBuilderException: PKIXCertPathBuilderImpl could not build a valid CertPath.; internal cause is: java.security.cert.CertPathValidatorException: The certificate issued by CN=DigiCert Global Root CA, OU= www.digicert.com<http://www.digicert.com >, O=DigiCert Inc, C=US is not trusted; internal cause is: java.security.cert.CertPathValidatorException: Certificate chaining error]   Solution   The java agent is using the IBM JVM's jre/lib/security/cacerts file to validate the controller's certificate. There are two ways to solve this problem: 1.) Use keytool to import the root of your controller's certificate chain, (the FTB CA's root cert), into jre/lib/security/cacerts 2.) Start your application with following JVM arguments: -Djavax.net.ssl.trustStore=/path/to/FTB_custom_trustStore.jks -Djavax.net.ssl.trustStorePassword=somepassword and make sure FTB_custom_trustStore.jks contains the FTB Certificate Authority root cert.   ... View more

Question Does the  JavaScript agent for EUEM (adrum.js) always need to be placed in the header of the web page? Won't this potentially cause latency and load delays in loading the web page? Answer Our Engineering team has gone to great lengths to ensure that the initial adrum.js is very lightweight.    Additionally, the loading of adrum.js is synchronous and minimal filesize to prevent page loading interruptions.  It calls out to the extension script, adrum-ext.js, asynchronously to allow rest of the page to load.   For most modern browsers, it does not matter where you copy adrum.js file on the page because they support navigation timing, which is what we use for the timings. So loading in the body for them is ok.   If you care about accurate timing for older browsers, mostly older IE and Safari, then you need to load the tag as early possible so the fall back timer can be started.    Using the adrum-start-time property gets us around the timing issues with older browsers. Place this JS snippet in the page's <head> and then place adrum.js at the bottom like so window["adrum-start-time"] = new Date().getTime(); More code examples here.   IE browsers tips: adrum.js loading placement in relation to <meta> tags definitely impacts pre-IE 9 browsers, therefore the recommended location for adrum.js is AFTER any <meta> tags With older browsers, load adrum.js as early as possible to avoid missing resource timing data Intranet settings in IE can prevent the JS agent from functioning properly   Browsers that support the navigation timing API can be found here: Browser Support for Navigation Timing API.   ... View more

Symptoms While trying to enable the End User Experience in the controller it returns an error similar to the following:   EUM license usage retrieval failed (Error fetching license usage information [org.apache.http.conn.HttpHostConnectException   For example, in an environment using the AppDynamics SaaS EUM the full error might read: EUM license usage retrieval failed (Error fetching license usage information [org.apache.http.conn.HttpHostConnectException: Connect to agg.eum-appdynamics.com:443 [agg.eum-appdynamics.com/52.10.146.82, agg.eum-appdynamics.com/54.186.213.46, agg.eum-appdynamics.com/54.148.146.12, agg.eum-appdynamics.com/54.191.252.35, agg.eum-appdynamics.com/54.213.8.211, agg.eum-appdynamics.com/54.69.27.161] failed: Connection refused while calling 'https://agg.eum-appdynamics.com/eumaggregator/licenseUsage'] )   Diagnosis To use EUM you must have port 443 open on the controller machine so the Controller can access the EUM cloud to collect EUM metrics. The EUM cloud can be the AppDynamics Cloud or your own if you have set up an on-premise EUM Server.   For  EUM in  AppDynamics Cloud Environment The Controller needs to access agg.eum-appdynamics.com through internet using HTTP over SSL on port :443.   For On-Premises EUM, the URL agg.eum-appdynamics.com will be different, but the issue is the same. Make sure port 443 is open.     Solution Confirm that port 443 is open and is not blocked by a firewall.   If you have a proxy in between the controller and your EUM cloud, configure the proxy as described in the following doc: Configure On-Prem Controller HTTP Proxy Settings.   Related documenation: Install and+Configure the On-Premise EUM Server. ... View more

This applies to versions  3.7 and later.   Use Case Description Using the flight search sample application, this example shows the use of the validation configuration. Prior to applying the custom correlation configuration with validation, the flow map shows the following.   Flow Map Without Correlation     Flow Map With Correlation and Validation Providing custom correlation configuration along with validation that specifies correlation for the requests for United Airlines searches, indicated by the value "UA232" in the getter chain configuration, we get the following flow map.     The validator configuration is specified in the validator-config element in addition to the standard producer and consumer XML shown in other examples. This XML snippet just shows the validator-config for this example. Place this element and its associated attributes and child elements inside the <intrumentation> element of your producer or consumer XML.   Validation XML Example Consumer XML configuration for custom-activity-correlation.xml <instrumentation> ... <validator-config> <validator class-match-type="MATCHES_CLASS" match-pattern="java.util.HashMap"> <data-match-config> <data-gatherer-type>POSITION</data-gatherer-type> <position>0</position> <getter-chain>this</getter-chain> <transformer-type>GETTER_METHODS</transformer-type> </data-match-config> </validator> <!-- no data match conditions with a class match validation --> <validator match-type="NOT_EMPTY"> <data-match-config> <data-gatherer-type>POSITION</data-gatherer-type> <position>0</position> <getter-chain>get(object/UA232)</getter-chain> <transformer-type>GETTER_METHODS</transformer-type> </data-match-config> </validator> </validator-config> </instrumentation> ... View more

Use Case Description SABRE has a Global Distribution System (GDS) that uses a custom protocol called SABRETalk. SABRETalk is a request-reply style protocol. Their subsidiary Travelocity would like to monitor their application with AppDyamics Pro. Java Agents will be installed on the GDS and Travelocity web servers. They want to see the correlation across these systems.   In this example reading and consuming the producer payload is handled by two different methods.   Correlation Example This example covers the following scenarios: Producer type - correlation method and exit point are the same Consumer type - correlation header read and continuing transaction are in different methods (downstream case) Producer Payload The producer payload being instrumented, Map<String, String> flights, comes from queryForStatus method. public Map<String, String> queryForStatus(Map<String, String> flights)   Producer Code SabreClient code snippet Call is issued to request the flight status information. .//Here we are sending the data using sockets... public Map<String, String> queryForStatus(Map<String, String> flights) throws IOException { ObjectOutputStream writer = null; OutputStream out = null; InputStream in = null; ObjectInputStream reader = null; Map<String, String> results = null;   Producer XML Configuration Producer XML configuration for custom-activity-correlation.xml This XML configuration tells AppDynamics how to instrument the exit point and where to add the correlation metadata. <!-- This configuration defines a custom exit and sets up the outgoing correlation downstream. On the consumer side the header is read and a transaction is begun at the same point. --> <activities> <producer> <!-- This instrumentation element defines a custom exit point. --> <instrumentation> <class-name>com.appdynamics.samples.FlightSearch</class-name> <method-name>queryForStatus</method-name> <match-type>MATCHES_CLASS</match-type> </instrumentation> <!-- This identifier element specifies how to name the custom exit point. This is similar to how we name custom exit points from the UI --> <identifiers> <identifier name="UnitedSearch"> <data-gatherer-type>INVOKED</data-gatherer-type> <getter-chain>this</getter-chain> <user-defined-name>UnitedSearch</user-defined-name> <transformer-type>USER_DEFINED_NAME</transformer-type> </identifier> </identifiers> <!-- The following correlation element specifies where and how to add the correlation metadata --> <correlation> <!-- This instrumentation element specifies how to get hold of the outgoing payload --> <instrumentation> <class-name>com.appdynamics.samples.FlightSearch</class-name> <method-name>queryForStatus</method-name> <match-type>MATCHES_CLASS</match-type> </instrumentation> <!-- This payload-pointer specifies where to grab the payload at the instrumentation point defined above --> <payload-pointer> <data-gatherer-type>POSITION</data-gatherer-type> <position>0</position> <getter-chain>this</getter-chain> <transformer-type>GETTER_METHODS</transformer-type> </payload-pointer> <!-- This payload-operation defines how to add the correlation metadata to the payload --> <payload-operation> <access-type>method</access-type> <access-method>put</access-method> <param-types>java.lang.Object, java.lang.Object</param-types> </payload-operation> </correlation> </producer> </activities>   Consumer Code Sabre Server code snippet The SABRETalk server is similar to the following where receiving and processing the request are handled by two separate methods. private Object readRequest(Socket client) { Object obj = null; try { ObjectInputStream in = new ObjectInputStream(client.getInputStream()); obj = in.readObject(); } catch (ClassCastException cce) { logger.log(Level.SEVERE, "Invalid Request", cce); } catch (ClassNotFoundException cnfe) { logger.log(Level.SEVERE, null, cnfe); } catch (IOException ex) { logger.log(Level.SEVERE, null, ex); } return obj; } private HashMap<String, Timestamp> processRequest(Map<String, String> payLoad) { return this.searcher.getFlightStatus(payLoad);   Consumer XML Configuration Consumer XML configuration for custom-activity-correlation.xml This XML configuration for the consumer specifies that this is a continuing transaction beginning at the specified entry point for the GDS tier and specifies where to find the correlation metadata in the message. <activities> <!-- The pojo-entry="true" specifies that the continuing transaction begins downstream. Note: The POJO continuing transaction begins only if correlation metadata was read upstream as defined by the <correlation> tag. If the correlation metadata could not be read then no downstream POJO transaction begins. --> <consumer pojo-entry="true"> <!-- This is where the correlation metadata is read and continuing transaction is begun --> <instrumentation> <class-name>com.appdynamics.samples.FlightStatusSearcher</class-name> <method-name>getFlightStatus</method-name> <match-type>MATCHES_CLASS</match-type> </instrumentation> <correlation> <instrumentation> <class-name>com.appdynamics.samples.FlightSearchHandler</class-name> <method-name>processRequest</method-name> <match-type>MATCHES_CLASS</match-type> </instrumentation> <!-- This is how we get the payload bearing our correlation metadata --> <payload-pointer> <data-gatherer-type>POSITION</data-gatherer-type> <position>0</position> <getter-chain>this</getter-chain> <transformer-type>GETTER_METHODS</transformer-type> </payload-pointer> <!-- This is how we extract correlation metadata from the payload --> <payload-operation> <access-type>method</access-type> <access-method>get</access-method> <param-types>java.lang.Object</param-types> </payload-operation> </correlation> </consumer> </activities>    Questions/Answers   Q: This example inserts the correlation key in the Map<String, String> flights map. W hen the correlation key is retrieved from the map: <payload-operation>  <access-type>method</access-type>  <access-method>get</access-method>  <param-types>java.lang.Object</param-types></payload-operation> Is it also removed from the map, or kept in the collection?   A: Once the correlation key is added to the map, it is kept there and not removed from it.   Q: Why is this java.lang.Object, java.lang.Object when the Map is <String,String>? Also, what is the signifance of this? Are you able to work with maps of other types? Integer Maps? Maps contain class Foo and class Bar key/values? What are the restrictions?   A: You cannot use your own classes when declaring the types for the map itself. This is because loading those classes is problematic as well as the correlation itself uses specific storage format. At this point, we support java.lang.Object or java.lang.String.     ... View more